This study uses student survey results to model retention. For the 2014 and 2015 cohorts of incoming freshmen, Term 3 enrollment was modelled based on predictor variables and their interactions. These terms were then applied in models where the data was subset into STEM majors and Non-STEM majors. Finally, the Term 3 enrollment of the 2017 class of incoming freshmen was predicted for all 3 groups. I found that Gender, Student of Color, Pell Grant Eligibility, whether or not St. Cloud State was the closest university to home, Belonging Index, Quality Points Predicted, and STEM were all important for the overall group. The interactions between Belonging and Gender, Belonging and Closest, Quality Points Predicted and Student of Color, and STEM and ACE were significant as well. The Non-STEM group was quite similar. For the STEM group however, the interactions weren’t as significant, and Gender was unimportant. For Term 3 enrollment of the 2017 class, the average predicted probability percentage was slightly higher than the enrollment percentage of the 2014-2015 classes.
Savage, Torrence, "Modeling Term 3 Retention Before Term 1 Completion" (2017). SCSU Data. 50.
Creative Commons License
This work is licensed under a Creative Commons Attribution-NonCommercial 4.0 International License